faceAI¶
Facial Keypoint Detection, via CNNs, Keras, and OpenCV!
The final Computer Vision project in the Artificial Intelligence Nanodegree program!
In this project,we’ll combine our knowledge of computer vision techniques and deep learning to build and end-to-end facial keypoint recognition system!
History¶
Facial keypoints are otable points around the eyes, nose, and mouth on any face and are used in many applications, from facial tracking to emotion recognition. There are three main parts to this project:
In this project you will get to explore a few of the many computer vision algorithms built into the OpenCV library. This expansive computer vision library is now almost 20 years old and still growing!
Setup¶
Part 1 : Investigating OpenCV, pre-processing, and face detection
- Step 0: Detect Faces Using a Haar Cascade Classifier
- Step 1: Add Eye Detection
- Step 2: De-noise an Image for Better Face Detection
- Step 3: Blur an Image and Perform Edge Detection
- Step 4: Automatically Hide the Identity of an Individual
Part 2 : Training a Convolutional Neural Network (CNN) to detect facial keypoints
- Step 5: Create a CNN to Recognize Facial Keypoints
- Step 6: Compile and Train the Model
- Step 7: Visualize the Loss and Answer Questions
Part 3 : Putting parts 1 and 2 together to identify facial keypoints on any image!
- Step 8: Build a Robust Facial Keypoints Detector (Complete the CV Pipeline)
Step 0: Detect Faces Using a Haar Cascade Classifier¶
Have you ever wondered how Facebook automatically tags images with your friends' faces with DeepFace? Or how high-end cameras automatically find and focus on a certain person's face?
Applications like these depend heavily on the machine learning task known as face detection - which is the task of automatically finding faces in images containing people.
At its root face detection is a classification problem - distinguishing between distinct classes of things. With face detection these distinct classes are 1) images of human faces and 2) everything else.
We use OpenCV's implementation of Haar feature-based cascade classifiers to detect human faces in images. OpenCV provides many pre-trained face detectors, stored as XML files on github!
We have downloaded one of these detectors and stored it in the detector_architectures directory.
Import Resources¶
In the next python cell, we load in the required libraries for this section of the project.
# Settings
%matplotlib inline
# Libraries
import numpy as np # Numpy for array management
import matplotlib.pyplot as plt # Matplotlib for plotting
import math # Math for...math?
import cv2 # OpenCV library for computer vision
from PIL import Image # Python Image Library to handle i
import time
Next, we load in and display a test image for performing face detection.
Note: by default OpenCV assumes the ordering of our image's color channels are Blue, then Green, then Red (BGR). This is slightly out of order with most image types whose color channels are ordered Red, then Green, then Blue (RGB).
In order to switch the Blue and Red channels of our test image around we will use OpenCV's cvtColor function, which you can read more about here; but in general, this utility function that can transform it for us, and do other transformations like converting a color image to grayscale, transforming a standard color image to HSV (Hue, Saturation, Lightness), etc.
# Load in color image for face detection
image = cv2.imread('images/test_image_1.jpg')
# Convert the image to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Plot our image using subplots to specify a size and title
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(1,1,1)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
There are a lot of people - and faces - in this picture. 13 faces to be exact!
In the next code cell, we demonstrate how to use a Haar Cascade classifier to detect all the faces in this test image.
This face detector uses information about patterns of intensity in an image to reliably detect faces under varying light conditions. So, to use this face detector, we'll first convert the image from color to grayscale.
But among these features we calculated, most of them are irrelevant. For example, consider the image above. Top row shows two good features:
1) The first feature selected seems to focus on the property that the region of the eyes is often darker than the region of the nose and cheeks.
2) The second feature selected relies on the property that the eyes are darker than the bridge of the nose.
But the same windows used on cheeks or any other place is irrelevant. So how do we select the best features out of 160000+ features? We brute force it with Adaboost!
We save a set of classifiers for each feature (eyes, nose, ears, etc. and make sure it can detect that featureset across images, then run that cascade of classifiers and see if we find faces! Here's someone smarter explaining it:
For this they introduced the concept of Cascade of Classifiers. Instead of applying all the 6000 features on a window, group the features into different stages of classifiers and apply one-by-one. (Normally first few stages will contain very less number of features). If a window fails the first stage, discard it. We don't consider remaining features on it. If it passes, apply the second stage of features and continue the process. The window which passes all stages is a face region. How is the plan !!!
Luckily, these guys even saved a fully trained architecture of the face detector -- found in the file haarcascade_frontalface_default.xml - and we can use it on our image to find faces! (To learn more about the parameters of the detector see this post.)
# Convert the RGB image to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Extract the pre-trained face detector from an xml file
# This Haar Cascade Classifier specializes in frontal face detection!
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image (Define bounding dimensions for the box)
faces = face_cascade.detectMultiScale(gray, 4, 6)
# Print the number of faces detected in the image
print('Number of faces detected:', len(faces))
# Make a copy of the orginal image to draw face detections on
image_with_detections = np.copy(image)
# Get the bounding box for each detected face
for (x,y,w,h) in faces:
# Add a red bounding box to the detections image
cv2.rectangle(image_with_detections, (x,y), (x+w,y+h), (255,0,0), 3)
# Display the image with the detections
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with Face Detections')
ax1.imshow(image_with_detections)
print(faces)
In the above code, faces is a numpy array, holding of detected faces! Awesome!
Here each row corresponds to a detected face, where each detected face is a 1D array with four entries that specifies the bounding box of the detected face:
x: The horizontal position of the detection box.y: The vertical position of the detection box.w: The width of the detection box.h: The height of the detection box.
If you think this depends a lot on the notion of this 'box', you're right! Shrinking or widening the shape of our bounding box, significanlty impacts our face detection ability. Let's see if we can improve our performance with even more features!
faces_smallbox = face_cascade.detectMultiScale(gray, 3, 3)
faces_mediumbox = face_cascade.detectMultiScale(gray, 5, 5)
faces_bigbox = face_cascade.detectMultiScale(gray, 7, 7)
# Get the bounding box for each detected face
def faceFinder(face_vector):
# Print the number of faces detected in the image
print('Number of faces detected:', len(face_vector))
# Make a copy of the orginal image to draw face detections on
image_with_detections = np.copy(image)
# Add a red bounding box to the detections image
[cv2.rectangle(image_with_detections, (x,y), (x+w,y+h), (255,0,0), 3) for (x,y,w,h) in face_vector]
# Display the image with the detections
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with Face Detections')
return ax1.imshow(image_with_detections)
faceFinder(faces_smallbox)
faceFinder(faces_mediumbox)
faceFinder(faces_bigbox)
Step 1: Add Eye Detections¶
Luckily, there are even more pre-trained detectors available that use a Haar Cascade Classifier - including full human body detectors, license plate detectors, and more! Since this strategy just depends on having cascades of valid classifiers, we can put a ton in a row and detect more!
A full list of the pre-trained architectures can be found here.
To test your eye detector, we'll first read in a new test image with just a single face.
# Load in color image for face detection
image = cv2.imread('images/james.jpg')
# Convert the image to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Plot the RGB image
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
Notice that even though the image is an original black and white image, we have to read it in as a color image that will still need to be converted to grayscale in order to perform the accurate face detection.
So, let's convert this image to grayscale, then load OpenCV's face detector and run it with parameters that detect this face accurately.
# Convert the image to Gray color space
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Plot the RGB image
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
# Extract the pre-trained face detector from an xml file
# This feature still uses frontal face detection
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image
faces = face_cascade.detectMultiScale(gray, 1.25, 6)
# Print the number of faces detected in the image
print('Number of faces detected:', len(faces))
# Make a copy of the orginal image to draw face detections on
image_with_detections = np.copy(image)
# Get the bounding box for each detected face
for (x,y,w,h) in faces:
# Add a red bounding box to the detections image
cv2.rectangle(image_with_detections, (x,y), (x+w,y+h), (255,0,0), 3)
# Display the image with the detections
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with Face Detection')
ax1.imshow(image_with_detections)
Add an eye detector to the current face detection setup.¶
let's build a Haar-cascade eye detector that can be included in the same way that the face detector was!
To set up an eye detector, we use the stored parameters of the eye cascade detector, called haarcascade_eye.xml, located in the detector_architectures subdirectory. Things to note:
First, let's make sure to give our loaded eye detector the variable name (
eye_cascade). We give the list of eye regions you detect the variable name (eyes).Second, since we've already run the face detector over this image, we should only search for eyes within the rectangular face regions detected in
faces(overlay classifiers!) This will minimize false detections, since there's a way to pre test it!
Lastly, once you've run your eye detector over the facial detection region, you should display the RGB image with both the face detection boxes (in red) and your eye detections (in green) to verify that everything works as expected.
# Make a copy of the original image to plot rectangle detections
image_with_detections = np.copy(image)
# Loop over the detections and draw their corresponding face detection boxes
for (x,y,w,h) in faces:
cv2.rectangle(image_with_detections, (x,y), (x+w,y+h),(255,0,0), 3)
# Do not change the code above this comment!
## Add eye detection, using haarcascade_eye.xml, to the current face detector algorithm
# Build on the existing cv2 object
eye_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_eye.xml')
eyes = eye_cascade.detectMultiScale(image_with_detections, 1.05, 4)
## Loop over the eye detections and draw their corresponding boxes in green on image_with_detections
for (x,y,w,h) in eyes:
cv2.rectangle(image_with_detections, (x,y), (x+w,y+h),(0,255,0), 2)
# Plot the image with both faces and eyes detected
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with Face and Eye Detection')
ax1.imshow(image_with_detections)
(Optional) Add face and eye detection to your laptop camera¶
It's time to kick it up a notch, and add face and eye detection to your laptop's camera! Afterwards, you'll be able to show off your creation like in the gif shown below - made with a completed version of the code!
Notice that not all of the detections here are perfect - and your result need not be perfect either. You should spend a small amount of time tuning the parameters of your detectors to get reasonable results, but don't hold out for perfection. If we wanted perfection we'd need to spend a ton of time tuning the parameters of each detector, cleaning up the input image frames, etc. You can think of this as more of a rapid prototype.
The next cell contains code for a wrapper function called laptop_camera_face_eye_detector that, when called, will activate your laptop's camera. You will place the relevant face and eye detection code in this wrapper function to implement face/eye detection and mark those detections on each image frame that your camera captures.
Before adding anything to the function, you can run it to get an idea of how it works - a small window should pop up showing you the live feed from your camera; you can press any key to close this window.
Note: Mac users may find that activating this function kills the kernel of their notebook every once in a while. If this happens to you, just restart your notebook's kernel, activate cell(s) containing any crucial import statements, and you'll be good to go!
### Add face and eye detection to this laptop camera function
# Make sure to draw out all faces/eyes found in each frame on the shown video feed
import cv2
import time
# wrapper function for face/eye detection with your laptop camera
def laptop_camera_go():
# Create instance of video capturer
cv2.namedWindow("face detection activated")
vc = cv2.VideoCapture(0)
# Try to get the first frame
if vc.isOpened():
rval, frame = vc.read()
else:
rval = False
# Keep the video stream open
while rval:
# Plot the image from camera with all the face and eye detections marked
cv2.imshow("face detection activated", frame)
# Exit functionality - press any key to exit laptop video
key = cv2.waitKey(20)
if key > 0: # Exit by pressing any key
# Destroy windows
cv2.destroyAllWindows()
# Make sure window closes on OSx
for i in range (1,5):
cv2.waitKey(1)
return
# Read next frame
time.sleep(0.05) # control framerate for computation - default 20 frames per sec
rval, frame = vc.read()
# Call the laptop camera face/eye detector function above
# laptop_camera_go()
Image quality is an important aspect of any computer vision task.
Typically, when creating a set of images to train a deep learning network, significant care is taken to ensure that training images are free of visual noise or artifacts that hinder object detection. While computer vision algorithms - like a face detector - are typically trained on 'nice' data such as this, new test data doesn't always look so nice!
Pre Processing¶
When applying a trained computer vision algorithm to a new piece of test data one often cleans it up first before feeding it in.
This sort of cleaning - referred to as pre-processing - can include a number of cleaning phases like blurring, de-noising, color transformations, etc., and many of these tasks can be accomplished using OpenCV.
In this short subsection we explore OpenCV's noise-removal functionality to see how we can clean up a noisy image, which we then feed into our trained face detector!
Create a noisy image to work with¶
In the next cell, we create an artificial noisy version of the previous multi-face image. This is a little exaggerated - we don't typically get images that are this noisy - but image noise, or 'grainy-ness' in a digitial image - is a fairly common phenomenon.
# Load in the multi-face test image again
image = cv2.imread('images/test_image_1.jpg')
# Convert the image copy to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Make an array copy of this image
image_with_noise = np.asarray(image)
# Create noise - here we add noise sampled randomly from a Gaussian distribution: a common model for noise
noise_level = 40
noise = np.random.randn(image.shape[0],image.shape[1],image.shape[2])*noise_level
# Add this noise to the array image copy
image_with_noise = image_with_noise + noise
# Convert back to uint8 format
image_with_noise = np.asarray([np.uint8(np.clip(i,0,255)) for i in image_with_noise])
# Plot our noisy image!
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Noisy Image')
ax1.imshow(image_with_noise)
In the context of face detection, the problem with an image like this is that - due to noise - we may miss some faces or get false detections.
In the next cell we apply the same trained OpenCV detector with the same settings as before. Let's see what sort of detections we get!
# Convert the RGB image to grayscale
gray_noise = cv2.cvtColor(image_with_noise, cv2.COLOR_RGB2GRAY)
# Extract the pre-trained face detector from an xml file
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image
faces = face_cascade.detectMultiScale(gray_noise, 4, 6)
# Print the number of faces detected in the image
print('Number of faces detected:', len(faces))
# Make a copy of the orginal image to draw face detections on
image_with_detections = np.copy(image_with_noise)
# Get the bounding box for each detected face
for (x,y,w,h) in faces:
# Add a red bounding box to the detections image
cv2.rectangle(image_with_detections, (x,y), (x+w,y+h), (255,0,0), 3)
# Display the image with the detections
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Noisy Image with Face Detections')
ax1.imshow(image_with_detections)
With this added noise we now miss one of the faces we had before!
Let's build in a pre-processing step to ensure we don't have this issue.
De-noise this image for better face detection¶
Time to get our hands dirty: using OpenCV's built in color image de-noising functionality called fastNlMeansDenoisingColored(), let's de-noise this image enough so that all the faces in the image are properly detected, and we can run it through our trained face detector.
What is noise? Noise is the arbitrary error of measurement. Using that notion, we can construct is a stochastic function, typically a random, normally distributed variable with zero mean.
For images, consider a noisy pixel, p = p_0 + n where p_0 is the true value of pixel and n is the noise in that pixel. You can take large number of same pixels (say N) from different images and computes their average. Ideally, you should get p = p_0, since mean of noise is zero, and over multiple samples the noisiness should cancel itself out
You can find its official documentation here and a useful example here, but here's the short version of what's happening:
The blue patches in the image looks the similar. Green patches looks similar. So we take a pixel, take small window around it, search for similar windows in the image, average all the windows and replace the pixel with the result we got. This method is Non-Local Means Denoising. It takes more time compared to blurring techniques we saw earlier, but its result is very good. More details and online demo can be found at first link in additional resources.
Note: you can keep all parameters except photo_render fixed as shown in the second link above. Play around with the value of this parameter - see how it affects the resulting cleaned image.
# %timeit!
## Use OpenCV's built in color image de-noising function to clean up our noisy image!
# fastNlMeansDenoisingColored(src[, dst[, h[, hColor[, templateWindowSize[, searchWindowSize]]]]])
denoised_image = cv2.fastNlMeansDenoisingColored(image_with_noise,None,19,19,7,21)
A notably expensive algorithm, but since it's using k-means clustering to generate consistent structure from noisy little windows, maybe it'll get our face detection back!
# Convert the RGB image to grayscale
gray_denoise = cv2.cvtColor(denoised_image, cv2.COLOR_RGB2GRAY)
# Extract the pre-trained face detector from an xml file
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image
faces = face_cascade.detectMultiScale(gray_denoise, 4, 6)
# Print the number of faces detected in the image
print('Number of faces detected:', len(faces))
# Make a copy of the orginal image to draw face detections on
denoised_image_with_detections = np.copy(denoised_image)
# Get the bounding box for each detected face
for (x,y,w,h) in faces:
# Add a red bounding box to the detections image
cv2.rectangle(denoised_image_with_detections, (x,y), (x+w,y+h), (255,0,0), 3)
# Display the image with the detections
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Denoised Image with Face Detections')
ax1.imshow(denoised_image_with_detections)
Now that we have developed a simple pipeline for detecting faces using OpenCV - let's start playing around with a few fun things we can do with all those detected faces!
Importance of Blur in Edge Detection¶
Edge detection is a concept that pops up almost everywhere in computer vision applications, as edge-based features (as well as features built on top of edges) are often some of the best features for e.g., object detection and recognition problems.
Edge detection is a dimension reduction technique - by keeping only the edges of an image we get to throw away a lot of non-discriminating information! This way, we only preserve the most important, global structures (ignoring local structures that aren't very discriminative).
As you can imagine, removing irrelevent local structures is a crucial pre-processing step to performing edge detection in an image, and it turns out blurring can do just that!
Below is an animated gif showing the result of an edge-detected cat taken from Wikipedia, where the image is gradually blurred more and more prior to edge detection.
When the animation begins you can't quite make out what it's a picture of, but as the animation evolves local structures are removed via blurring the cat becomes visible in the edge-detected image.
Edge detection is a convolution performed on the image itself, and you can read about Canny edge detection on this OpenCV documentation page.
Canny edge detection¶
In the cell below we load in a test image, then apply Canny edge detection on it. The original image is on the left, and the edge-detected version is on the right.
Notice how the result looks very busy - there are too many little details preserved in the image before it is sent to the edge detector!
When applied in computer vision applications, edge detection should preserve global structure; doing away with local structures that don't help describe what objects are in the image.
# Load in the image
image = cv2.imread('images/fawzia.jpg')
# Convert to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Convert to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Perform Canny edge detection
edges = cv2.Canny(gray,100,200)
# Dilate the image to amplify edges
edges = cv2.dilate(edges, None)
# Plot the RGB and edge-detected image
fig = plt.figure(figsize = (15,15))
ax1 = fig.add_subplot(121)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
ax2 = fig.add_subplot(122)
ax2.set_xticks([])
ax2.set_yticks([])
ax2.set_title('Canny Edges')
ax2.imshow(edges, cmap='gray')
Without first blurring the image, and removing small, local structures, a lot of irrelevant edge content gets picked up and amplified by the detector (as shown in the right panel above).
Blur the image then perform edge detection¶
So, instead of all these fine details, lets blur our image and figure out we're generally looking at, i.e. dropping local noise for global structure!
In the next cell, we will repeat this experiment - blurring the image first to remove these local structures, so that only the important boudnary details remain in the edge-detected image.
Lets blur the image by using OpenCV's filter2d() functionality - which is discussed in this documentation page - and use an averaging kernel of width equal to 4.
So what's going on within the function? In general, convolution is an operation between every part of an image and an operator (kernel); using linear algebra to transform image matrices!
A kernel is essentially a fixed size array of numerical coefficeints along with an anchor point in that array, which is tipically located at the center.
Therefore, the value of a convolution is calculated in the following way:
1) Place the kernel anchor on top of a determined pixel, with the rest of the kernel overlaying the corresponding local pixels in the image. 2) Multiply the kernel coefficients by the corresponding image pixel values and sum the result. 3) Place the result to the location of the anchor in the input image. 4) Repeat for all pixels by scanning the kernel over the entire image.
Expressing the procedure above in the form of an equation we would have:
$ H(x,y) = \sum_{i=0}^{M_{i} - 1} \sum_{j=0}^{M_{j}-1} I(x+i - a_{i}, y + j - a_{j})K(i,j) $
Fortunately, OpenCV provides you with the function filter2D() so you do not have to code all these operations!
### Blur the test imageusing OpenCV's filter2d functionality,
# Load in the image
image = cv2.imread('images/fawzia.jpg')
# Convert to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Convert to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Use an averaging kernel, and a kernel width equal to 4
kernel = np.ones((4,4),np.float32)/16
blurred = cv2.filter2D(gray,-1,kernel)
# Perform Canny edge detection
edges = cv2.Canny(blurred,100,200)
# Dilate the image to amplify edges
edges = cv2.dilate(edges, None)
# Plot the RGB and edge-detected image
fig = plt.figure(figsize = (15,15))
ax1 = fig.add_subplot(121)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
ax2 = fig.add_subplot(122)
ax2.set_xticks([])
ax2.set_yticks([])
ax2.set_title('Canny Edges')
ax2.imshow(edges, cmap='gray')
Step 4: Automatically Hide the Identity of an Individual¶
If you film something like a documentary or reality TV, you must get permission from every individual shown on film before you can show their face, otherwise you need to blur it out - by blurring the face a lot (so much so that even the global structures are obscured)! This is also true for projects like Google's StreetView maps - an enormous collection of mapping images taken from a fleet of Google vehicles. Because it would be impossible for Google to get the permission of every single person accidentally captured in one of these images they blur out everyone's faces, the detected images must automatically blur the identity of detected people. Here's a few examples of folks caught in the camera of a Google street view vehicle.
Read in an image to perform identity detection¶
Let's try this out for ourselves. Use the face detection pipeline built above and what you know about using the filter2D to blur and image, and use these in tandem to hide the identity of the person in the following image - loaded in and printed in the next cell.
# Load in the image
image = cv2.imread('images/gus.jpg')
# Convert the image to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Display the image
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
Use blurring to hide the identity of an individual in an image¶
Let's identify a face, then blur it out!
Our code should:
1) Automatically detect the face in this image.
2) Blur it out!
Make sure to adjust the parameters of the averaging blur filter to completely obscure this person's identity.
#Implement face detection
image_with_blur_face = np.copy(image)#saved to used later after we detect the faces
#image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image
faces = face_cascade.detectMultiScale(gray, 1.13, 6)
# Using averaging blur filter
kernel = np.ones((80, 80),np.float32) / 6400
## Blur the bounding box around each detected face using an averaging filter and display the result
for (x,y,w,h) in faces:
# Too small - filtering mistakes; scale up!
if(w < 100) | (h < 100):
continue
# Blur bounding box to the detections image
image_with_blur_face[y:y+h, x:x+w] = cv2.filter2D(image_with_blur_face[y:y+h, x:x+w], -1, kernel)
# Display the image
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with blur face')
ax1.imshow(image_with_blur_face)
(Optional) Build identity protection into your laptop camera¶
In this optional task you can add identity protection to your laptop camera, using the previously completed code where you added face detection to your laptop camera - and the task above. You should be able to get reasonable results with little parameter tuning - like the one shown in the gif below.
As with the previous video task, to make this perfect would require significant effort - so don't strive for perfection here, strive for reasonable quality.
The next cell contains code a wrapper function called laptop_camera_identity_hider that - when called - will activate your laptop's camera. You need to place the relevant face detection and blurring code developed above in this function in order to blur faces entering your laptop camera's field of view.
Before adding anything to the function you can call it to get a hang of how it works - a small window will pop up showing you the live feed from your camera, you can press any key to close this window.
Note: Mac users may find that activating this function kills the kernel of their notebook every once in a while. If this happens to you, just restart your notebook's kernel, activate cell(s) containing any crucial import statements, and you'll be good to go!
### Insert face detection and blurring code into the wrapper below to create an identity protector on your laptop!
import cv2
import time
def laptop_camera_go():
# Create instance of video capturer
cv2.namedWindow("face detection activated")
vc = cv2.VideoCapture(0)
# Try to get the first frame
if vc.isOpened():
rval, frame = vc.read()
else:
rval = False
# Keep video stream open
while rval:
# Plot image from camera with detections marked
cv2.imshow("face detection activated", frame)
# Exit functionality - press any key to exit laptop video
key = cv2.waitKey(20)
if key > 0: # Exit by pressing any key
# Destroy windows
cv2.destroyAllWindows()
for i in range (1,5):
cv2.waitKey(1)
return
# Read next frame
time.sleep(0.05) # control framerate for computation - default 20 frames per sec
rval, frame = vc.read()
# Run laptop identity hider
#laptop_camera_go()
Step 5: Create a CNN to Recognize Facial Keypoints¶
Alright, we're ready to recognize faces!
In practice, the OpenCV Library is often used with other machine learning and deep learning libraries to produce interesting results.
In this stage of the project we will create our own end-to-end pipeline - employing convolutional networks in keras along with OpenCV - to apply a "selfie" filter to streaming video and images.
We will start by creating and then training a convolutional network that can detect facial keypoints in a small dataset of cropped images of human faces. Then, we'll use OpenCV to expand our detection algorithm to more general images.
First things first, What are facial keypoints? Let's take a look at some examples.
Facial keypoints (also called facial landmarks) are the small blue-green dots shown on each of the faces in the image above - there are 15 keypoints marked in each image. They mark important areas of the face - the eyes, corners of the mouth, the nose, etc.
Facial keypoints can be used in a variety of machine learning applications from face and emotion recognition to commercial applications like the image filters popularized by Snapchat.
Below we illustrate a filter that, using the results of this section, automatically places sunglasses on people in images (using the facial keypoints to place the glasses correctly on each face). Here, the facial keypoints have been colored lime green for visualization purposes:
Make a facial keypoint detector¶
So, how can we make a facial keypoint detector?
Well, at a high level, notice that facial keypoint detection is a regression problem: A single face corresponds to a set of 15 facial keypoints (a set of 15 corresponding $(x, y)$ coordinates, i.e., an output point).
Because our input data are images, we can employ a convolutional neural network to recognize patterns in our images and learn how to identify these keypoint given sets of labeled data.
In order to train a regressor, we need a training set - a set of facial image / facial keypoint pairs to train on. For this we will be using this dataset from Kaggle.
We've already downloaded this data and placed it in the data directory. Make sure that you have both the training and test data files!
The training dataset contains several thousand $96 \times 96$ grayscale images of cropped human faces, along with each face's 15 corresponding facial keypoints (also called landmarks) that have been placed by hand (!), and recorded in $(x, y)$ coordinates.
This wonderful resource also has a substantial testing set, which we will use in tinkering with our convolutional network!
Load Data¶
To load in this data, run the Python cell below - notice we will load in both the training and testing sets.
The load_data() function is in the included in the handy utils.py file.
from utils import *
# Load training set
X_train, y_train = load_data()
print("X_train.shape == {}".format(X_train.shape))
print("y_train.shape == {}; y_train.min == {:.3f}; y_train.max == {:.3f}".format(
y_train.shape, y_train.min(), y_train.max()))
# Load testing set
X_test, _ = load_data(test=True)
print("X_test.shape == {}".format(X_test.shape))
The load_data() function in utils.py originates from this excellent blog post, which you are strongly encouraged to read.
def load_data(test=False):
"""
Loads data from FTEST if *test* is True, otherwise from FTRAIN.
Important that the files are in a `data` directory
"""
FTRAIN = 'data/training.csv'
FTEST = 'data/test.csv'
# Testing data or training data?
fname = FTEST if test else FTRAIN
# Read it's csv into a pandas DataFrame
df = read_csv(os.path.expanduser(fname))
# The Image column has pixel values separated by space; convert to numpy array:
df['Image'] = df['Image'].apply(lambda im: np.fromstring(im, sep=' '))
# Drop all rows that have missing values in them
df = df.dropna()
# Normalize pixel values to [0, 1] via scaling
X = np.vstack(df['Image'].values) / 255.
X = X.astype(np.float32)
# Reshape back into images as 96 x 96 x 1
X = X.reshape(-1, 96, 96, 1)
# only FTRAIN has target columns
if not test:
y = df[df.columns[:-1]].values # target coordinates
y = (y - 48) / 48 # scale to [-1, 1]
X, y = shuffle(X, y, random_state=42) # shuffle train data
y = y.astype(np.float32) # return as float32
else:
y = None
return X, y
Note how the output values - that is, the coordinates of each set of facial landmarks - have been normalized to take on values in the range $[-1, 1]$, while the pixel values of each input point (a facial image) have been normalized to the range $[0,1]$.
Note: the original Kaggle dataset contains some images with several missing keypoints. For simplicity, the load_data() function removes those images with missing labels from the dataset. As an optional extension, you are welcome to amend the load_data function to include the incomplete data points!
Visualize the Training Data¶
Execute the code cell below to visualize a subset of the training data.
import matplotlib.pyplot as plt
%matplotlib inline
fig = plt.figure(figsize=(20,20))
fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)
for i in range(9):
ax = fig.add_subplot(3, 3, i + 1, xticks=[], yticks=[])
plot_data(X_train[i], y_train[i], ax)
For each training image, there are:
- 2 landmarks per eyebrow (4 total).
- 3 per eye (6 total).
- 4 for the mouth.
- 1 for the tip of the nose.
Review the plot_data() function in utils.py to understand how the 30-dimensional training labels in y_train are mapped to facial locations, as this function will prove useful for your pipeline!
def plot_data(img, landmarks, axis):
"""
Plot image (img), along with normalized facial keypoints (landmarks)
"""
# plot the image
axis.imshow(np.squeeze(img), cmap='gray')
# undo the normalization (!?)
landmarks = landmarks * 48 + 48
# Plot the keypoints
axis.scatter(landmarks[0::2],
landmarks[1::2],
marker='o',
c='c',
s=40)
Specify the CNN Architecture¶
In this section, you will specify a neural network for predicting the locations of facial keypoints. Use the code cell below to specify the architecture of your neural network. We have imported some layers that you may find useful for this task, but if you need to use more Keras layers, feel free to import them in the cell.
Your network should accept a $96 \times 96$ grayscale image as input, and it should output a vector with 30 entries, corresponding to the predicted (horizontal and vertical) locations of 15 facial keypoints. If you are not sure where to start, you can find some useful starting architectures in this blog, but you are not permitted to copy any of the architectures that you find online.
Intution¶
We are allowed to talk about this awesome example though! Let's see what's going on.
1) Striding: Stride controls how the kernel / filter moves around the input shape. It's basically the number of squares we can move by when we downsample; notice the output for a Stride of 1 vs. a Stride of 2: (number of squares we can move by),
2) Padding: As we can see, the output field for a larger stride is smaller the bigger we step; sooner or later, we'd have issues fitting our kernel onto the squashed images. That's where padding can help; if we set our padding to size 2, we add a little buffer room about the image that results in a (36 x 36 x 2) shape throughought our network, even while we downsample!
3) Convolution: Therefore, we can describe this convolutional layer in a mathematical space. Let's use O is the output height/length, W is the input height/length, K is the filter size, P is the padding, and S is the stride:
4) Activation: Researchers like Dr. Hinton have discovered out that ReLU layers work far better then most other activators, especially with achieving computational efficiency without making a significant difference to the accuracy. It also helps to alleviate the vanishing gradient problem, which is the issue where the lower layers of the network train very slowly because the gradient decreases exponentially through the layers. ReLU activation just applies the function $f(x) = max(0, x)$ to all of the values in the input volume, so we get 0, an absolute integer, and no negatives.
5) Pooling: Downsampling post activation! Unlike a typical convolution, we are picking standard 'pool' sizes with matching strides and padding, and will not be generating any trainable parameters that can be affected via ReLU activators and backpropogation. All we're getting a smaller / zoomed in image to work with; and feed into the next step!
As we convolve, pool, dropout, and repeat, we can train our net to recognize the features we're looking for; and where to put facial keypoints! Later, we'll even compare our performance using the Haar Caascade classifier (and smush 'em together!)
# Import deep learning resources from Keras
from keras.models import Sequential
from keras.layers import Convolution2D, MaxPooling2D, Dropout, GlobalAveragePooling2D, Flatten, Dense
## A Convolutional Neural Network that:
# * Accept 96x96 pixel graysale images as input.
# * Use a fully-connected convolutional neural network to classify facial features.
# * Output a classification for 30 values (2 for each facial keypoint)
model = Sequential() # A Linear CNN Layer Stack
model.add(Convolution2D(filters=32,
kernel_size=3,
padding='valid',
activation='relu',
input_shape=X_train.shape[1:])) # Layer 1: 2D Convolution,
# 32 Filters (independent kernels)
# Using a 3x3 Kernel for the operation,
# 'Same' style padding for uniform image size
# Rectified Linear Unit activation,
# Shape is the input layer, so use first sample shape
model.add(MaxPooling2D(pool_size=2)) # Layer 2: Pooling
# Pool / Average input using 2x2 kernel
# Reduce size of kernel (and resulting computation)
# Let's use increase filters and avoid overfitting
model.add(Dropout(0.2)) # Layer 3: Dropout
# Drop paramaters to avoid overfitting
# Somehow even improves performance. Dope.
model.add(Convolution2D(filters=64,
kernel_size=2,
padding='same',
activation='relu')) # Layer 4: 2D Convolution,
# 128 Filters at 2x2, same padding, ReLU activation
model.add(MaxPooling2D(pool_size=2)) # Layer 5: Pooling
# Pool / average input using 2x2 kernel
model.add(Dropout(0.2)) # Layer 6: Dropout
# Drop 20% of paramaters
model.add(Convolution2D(filters=128,
kernel_size=2,
padding='same',
activation='relu')) # Layer 7: 2D Convolution,
# 128 Filters at 2x2, same padding, ReLU activation
model.add(MaxPooling2D(pool_size=2)) # Layer 8: Pooling
# Pool / average input using 2x2 kernel
model.add(Dropout(0.2)) # Layer 9: Dropout
# Drop 20% of paramaters
model.add(Convolution2D(filters=256,
kernel_size=2,
padding='same',
activation='relu')) # Layer 10: 2D Convolution,
# 256 Filters at 2x2, same padding, ReLU activation
model.add(MaxPooling2D(pool_size=2)) # Layer 11: Pooling
# Pool / average input using 2x2 kernel
model.add(Dropout(0.2)) # Layer 12: Dropout
# Drop 20% of paramaters
model.add(Convolution2D(filters=512,
kernel_size=2,
padding='same',
activation='relu')) # Layer 13: 2D Convolution,
# 512 Filters! at 2x2, same padding, ReLU activation
model.add(MaxPooling2D(pool_size=2)) # Layer 11: Pooling
# Pool / average input using 2x2 kernel
model.add(Dropout(0.2)) # Layer 12: Dropout
# Drop 20% of paramaters
model.add(GlobalAveragePooling2D()) # Layer 13: Global 2D Pooling
# Global pooling to ensure structural regularization
# Evidently good for avoiding overfitting
model.add(Dense(30)) # Layer 14: Dense Layer
# Dense layer for 30 categories / 15 keypoint features
model.summary() # Model Summary
Step 6: Compile and Train the Model¶
After specifying your architecture, you'll need to compile and train the model to detect facial keypoints'
Compile and Train the Model¶
Use the compile() method to configure the learning process. Be sure to experiment with your choice of optimizer; you may have some ideas about which will work best (SGD vs. RMSprop, etc), but take the time to empirically verify your theories.
Use the fit() method to train the model. Break off a validation set by setting validation_split=0.2. Save the returned History object in the history variable.
Experiment with your model to minimize the validation loss (measured as mean squared error). A very good model will achieve about 0.0015 loss (though it's possible to do even better). When you have finished training, save your model as an HDF5 file with file path my_model.h5.
from keras.optimizers import SGD, RMSprop, Adagrad, Adadelta, Adam, Adamax, Nadam
from keras.callbacks import ModelCheckpoint
## Compile the model
model.compile(optimizer='Adamax', loss="mean_squared_error", metrics=['accuracy'])
## Save the best model as model.h5 (!)
checkpointer = ModelCheckpoint(filepath='my_model.h5',
verbose=1, save_best_only=True)
## Train the model
hist = model.fit(X_train, y_train, batch_size=32, epochs=128, verbose=1,
validation_split=0.2, callbacks=[checkpointer])
Quick reference point when you're attempting to run the code above; I had to leave this computer on for at least half a day to roll the loss and accuracy to these levels. I literally had to restart the kernel to get my mode!
The good news is the save() function in our epoch step ensures we always save the best available model. This is where stuff like TensorFlow and Keras really shine; multiple layers of well thought out abstractions (in an occasionally opinionated syntax.)
from keras.models import load_model
# Loading the best model found
model = load_model('my_model.h5')
Answer a few questions and visualize the loss¶
Question 1: Outline the steps you took to get to your final neural network architecture and your reasoning at each step.
Answer: This structure is a fairly standard CNN skeleton, lifted and retrofit from some earlier work. I'd like to discuss the convolution(), pooling(), dropout() substack, because it's role seems crucial to our performance and explored much more thoroughly in this paper). To poorly snynthesize, I feel it to give us 4 things:
- Scaling Filters: Without pooling and dropout, the number of trainable parameters would get out of control too fast; we're already at 712K trainable values, and thats already with 20% decrease in connectivity every stage.
- Kernel Pooling: Reducing images to smaller, denser vectors seems counterintuitive, but the intuition seems very similar to the Gaussian blurring we used for edge detection earlier; by blurring or otherwise transforming a region, we drop local structure for global structures.
- Cascading Convolution: Just like the earlier cascade filter, we are best served by analyzing the existence of 'feature groups', which is kind of what this CNN is stumbling upon, especially by using the well labelled keypoint data to correct via backprop. This idealogy has been further evolved using Capsule Networks from the OG himself, George Hinton.
- Dropout: Dropout is a remarkably useful tool for preventing overfitting. I have still to understand why fewer connections yield such better results, though a little noise seems to great improve the performance. For some reason or another, gradient descent goes smoother on slightly rougher, unknown terrain.
After all those layers I we added a globalaveragepool() of ResNet-50 fame. It aims to enforce global feature detection, further preventing overfitting, before a final dense() layer to classify our feature detection output via 30 densly connected linear classifiers.
Question 2: Defend your choice of optimizer. Which optimizers did you test, and how did you determine which worked best?
Answer: The Adamax is an evolution of the Adam optimizer, which computes individual adaptive learning rates for different parameters from estimates of first and second moments of the gradients; hence it's name, Adaptive Moment estimation. It sounds like just bias-free RMSProp estimation, which is more cost effective. Adamax is a further expansion of the Adam optimizer, whose update rule for individual weights is to scale their gradients inversely proportional to a (scaled) $L^2$ norm of their current and past gradients. If we generalize this $L^2$ norm-based update rule to an $L^p$ norm based update rule, we run into variants that become numerically unstable for large $p$. However, a special case emerges for just the limit, i.e. $p → ∞$; a surprisingly simple and stable algorithm emerges! Frankly speaking, some of the math past this point is well beyond me, but as a new and interesting optimizer I was eager to test it in the field.
Use the code cell below to plot the training and validation loss of your neural network. You may find this resource useful.
## Visualize the training and validation loss of your neural network
# Available historical data
# print(hist.history.keys())
# Accuracy History
plt.plot(hist.history['acc'])
plt.plot(hist.history['val_acc'])
plt.title('Model Accuracy')
plt.ylabel('accuracy')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()
# Loss History
plt.plot(hist.history['loss'])
plt.plot(hist.history['val_loss'])
plt.title('Model Loss')
plt.ylabel('loss')
plt.xlabel('epoch')
plt.legend(['train', 'test'], loc='upper left')
plt.show()
Question 3: Do you notice any evidence of overfitting or underfitting in the above plot? If so, what steps have you taken to improve your model? Note that slight overfitting or underfitting will not hurt your chances of a successful submission, as long as you have attempted some solutions towards improving your model (such as regularization, dropout, increased/decreased number of layers, etc).
Answer:
In past projects with fewer layers and no dropout, I'd achieve some pretty horific losses. I'd hope that the more conservative, dropout() intensive architecture with the new Adamax optimizer should yield much better results. I also saw an increase in performance when I used that sub-network 4 times instead of just 3. I'd be open to any suggestions! Also, earlier runs of the training had a very close train and test loss, but it seems here that were are beginning to overfit quite severely to our training data. At least we're only talking about loss around 0.0010.
Visualize a Subset of the Test Predictions¶
Execute the code cell below to visualize your model's predicted keypoints on a subset of the testing images.
y_test = model.predict(X_test)
fig = plt.figure(figsize=(20,20))
fig.subplots_adjust(left=0, right=1, bottom=0, top=1, hspace=0.05, wspace=0.05)
for i in range(9):
ax = fig.add_subplot(3, 3, i + 1, xticks=[], yticks=[])
plot_data(X_test[i], y_test[i], ax)
Step 8: Complete the pipeline¶
With the work you did in Sections 1 and 2 of this notebook, along with your freshly trained facial keypoint detector, you can now complete the full pipeline. That is given a color image containing a person or persons you can now
- Detect the faces in this image automatically using OpenCV
- Predict the facial keypoints in each face detected in the image
- Paint predicted keypoints on each face detected
In this Subsection you will do just this!
Facial Keypoints Detector¶
Finally, let's use the OpenCV face detection functionality we built in previous Sections to expand the functionality of your keypoints detector to color images with arbitrary size. Our function should perform the following steps:
- Accept a color image.
- Convert the image to grayscale.
- Detect and crop the face contained in the image.
- Locate the facial keypoints in the cropped image.
- Overlay the facial keypoints in the original (color, uncropped) image.
Note: step 4 can be the trickiest because remember your convolutional network is only trained to detect facial keypoints in $96 \times 96$ grayscale images where each pixel was normalized to lie in the interval $[0,1]$, and remember that each facial keypoint was normalized during training to the interval $[-1,1]$. This means - practically speaking - to paint detected keypoints onto a test face you need to perform this same pre-processing to your candidate face - that is after detecting it you should resize it to $96 \times 96$ and normalize its values before feeding it into your facial keypoint detector. To be shown correctly on the original image the output keypoints from your detector then need to be shifted and re-normalized from the interval $[-1,1]$ to the width and height of your detected face.
When complete you should be able to produce example images like the one below
# Load in color image for face detection
image = cv2.imread('images/obamas4.jpg')
# Convert the image to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# plot our image
fig = plt.figure(figsize = (9,9))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('image copy')
ax1.imshow(image)
from keras.models import load_model
model = load_model('my_model.h5')
### Using the face detection code from Section 1 with our trained neural net:
## * Load the model using the best discovered weights (saved to my_model.h5).
## * Recognize faces, then predict where keypoints should go!
## * Paint those predicted keypoints on the test image!
def find_keypoints(image):
# Convert the RGB image to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)
# Extract the pre-trained face detector from an xml file
face_cascade = cv2.CascadeClassifier('detector_architectures/haarcascade_frontalface_default.xml')
# Detect the faces in image
faces = face_cascade.detectMultiScale(gray, 1.25, 6)
# Print the number of faces detected in the image
print('Number of faces detected:', len(faces))
# Make a copy of the orginal image to draw face detections on
image_with_predicted_keypoints = np.copy(image)
faces_keypoints = []
for (x,y,w,h) in faces:
# Crop and resize the face to 96 x 96
face_image = gray[y:y+h , x:x+w]
resize_face = cv2.resize(face_image, (96, 96))
# Normalize and convert to the input format
normalized_face = resize_face / 255
normalized_face = normalized_face[np.newaxis, :, :, np.newaxis]
# Predict the face keypoints and un-normalized the coordinates
keypoints = model.predict(normalized_face)
# Just tried the voodoo from earlier; why is 48 (96/2) untransform the normalization?
keypoints = keypoints * 48 + 48
# Convert keypoints to (x,y)
x_coordinate, y_coordinate = keypoints[0][0::2], keypoints[0][1::2]
# Scale x, y back up to
x_coordinate, y_coordinate = x + x_coordinate * w / 96, y + y_coordinate * h / 96
# Use w, h to draw red boxes around the coordinates
cv2.rectangle(image_with_predicted_keypoints, (x,y), (x+w,y+h), (255,0,0), 3)
# Draw green circles around the keypoints!
faces_keypoints.append((x_coordinate, y_coordinate))
for xc, yc in zip(x_coordinate, y_coordinate):
cv2.circle(image_with_predicted_keypoints, (xc, yc), 1, (0, 255, 0), 3)
return image_with_predicted_keypoints, faces_keypoints
# Run Function
image_with_predicted_keypoints, faces_keypoints = find_keypoints(image)
# Display the image with the keypoints
fig = plt.figure(figsize = (9,9))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Image with predicted keypoints')
ax1.imshow(image_with_predicted_keypoints)
Further Directions - add a filter using facial keypoints to your laptop camera¶
Now you can add facial keypoint detection to your laptop camera - as illustrated in the gif below.
The next Python cell contains the basic laptop video camera function used in the previous optional video exercises. Combine it with the functionality you developed for keypoint detection and marking in the previous exercise and you should be good to go!
import cv2
import time
from keras.models import load_model
def laptop_camera_go():
# Create instance of video capturer
cv2.namedWindow("face detection activated")
vc = cv2.VideoCapture(0)
# Try to get the first frame
if vc.isOpened():
rval, frame = vc.read()
else:
rval = False
# keep video stream open
while rval:
# plot image from camera with detections marked
cv2.imshow("face detection activated", frame)
# exit functionality - press any key to exit laptop video
key = cv2.waitKey(20)
if key > 0: # exit by pressing any key
# destroy windows
cv2.destroyAllWindows()
# hack from stack overflow for making sure window closes on osx --> https://stackoverflow.com/questions/6116564/destroywindow-does-not-close-window-on-mac-using-python-and-opencv
for i in range (1,5):
cv2.waitKey(1)
return
# read next frame
time.sleep(0.05) # control framerate for computation - default 20 frames per sec
rval, frame = vc.read()
# Run your keypoint face painter
# laptop_camera_go()
(Optional) Further Directions - add a filter using facial keypoints¶
Using your freshly minted facial keypoint detector pipeline you can now do things like add fun filters to a person's face automatically. In this optional exercise you can play around with adding sunglasses automatically to each individual's face in an image as shown in a demonstration image below.
To produce this effect an image of a pair of sunglasses shown in the Python cell below.
# Load in sunglasses image - note the usage of the special option
# cv2.IMREAD_UNCHANGED, this option is used because the sunglasses
# image has a 4th channel that allows us to control how transparent each pixel in the image is
sunglasses = cv2.imread("images/sunglasses_4.png", cv2.IMREAD_UNCHANGED)
# Plot the image
fig = plt.figure(figsize = (6,6))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.imshow(sunglasses)
ax1.axis('off');
This image is placed over each individual's face using the detected eye points to determine the location of the sunglasses, and eyebrow points to determine the size that the sunglasses should be for each person (one could also use the nose point to determine this).
Notice that this image actually has 4 channels, not just 3.
# Print out the shape of the sunglasses image
print ('The sunglasses image has shape: ' + str(np.shape(sunglasses)))
It has the usual red, blue, and green channels any color image has, and a *4th channel representing the transparency level of each pixel in the image8.
Here's how the transparency channel works: the lower the value, the more transparent the pixel will become. The lower bound (completely transparent) is zero here, so any pixels set to 0 will not be seen.
This is how we can place this image of sunglasses on someone's face and still see the area around of their face where the sunglasses lie - because these extra pixels in the sunglasses image have been made completely transparent!
Lets check out the alpha channel of our sunglasses image in the next Python cell. Note because many of the pixels near the boundary are transparent we'll need to explicitly print out non-zero values if we want to see them.
# Print out the sunglasses transparency (alpha) channel
alpha_channel = sunglasses[:,:,3]
print ('the alpha channel here looks like')
print (alpha_channel)
# Just to double check that there are indeed non-zero values
# Let's find and print out every value greater than zero
values = np.where(alpha_channel != 0)
print ('\n the non-zero values of the alpha channel look like')
print (values)
This means that when we place this sunglasses image on top of another image, we can use the transparency channel as a filter to tell us which pixels to overlay on a new image (only the non-transparent ones with values greater than zero).
One last thing: it's helpful to understand which keypoint belongs to the eyes, mouth, etc. So, in the image below, we also display the index of each facial keypoint directly on the image so that you can tell which keypoints are for the eyes, eyebrows, etc.
With this information, we're well on our way to completing this filtering task! Let's see if you can place the sunglasses automatically on the individuals in the image loaded in / shown in the next Python cell.
# Load in color image for face detection
image = cv2.imread('images/obamas4.jpg')
# Convert the image to RGB colorspace
image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
# Plot the image
fig = plt.figure(figsize = (8,8))
ax1 = fig.add_subplot(111)
ax1.set_xticks([])
ax1.set_yticks([])
ax1.set_title('Original Image')
ax1.imshow(image)
## (Optional) Use the face detection code we saw in Section 1 with your trained conv-net to put
## sunglasses on the individuals in our test image
Further Directions - add a filter using facial keypoints to your laptop camera¶
Now you can add the sunglasses filter to your laptop camera - as illustrated in the gif below.
The next Python cell contains the basic laptop video camera function used in the previous optional video exercises. Combine it with the functionality you developed for adding sunglasses to someone's face in the previous optional exercise and you should be good to go!
import cv2
import time
from keras.models import load_model
import numpy as np
def laptop_camera_go():
# Create instance of video capturer
cv2.namedWindow("face detection activated")
vc = cv2.VideoCapture(0)
# try to get the first frame
if vc.isOpened():
rval, frame = vc.read()
else:
rval = False
# Keep video stream open
while rval:
# Plot image from camera with detections marked
cv2.imshow("face detection activated", frame)
# Exit functionality - press any key to exit laptop video
key = cv2.waitKey(20)
if key > 0: # exit by pressing any key
# Destroy windows
cv2.destroyAllWindows()
for i in range (1,5):
cv2.waitKey(1)
return
# Read next frame
time.sleep(0.05) # control framerate for computation - default 20 frames per sec
rval, frame = vc.read()
# Load facial landmark detector model
# model = load_model('my_model.h5')
# Run sunglasses painter
# laptop_camera_go()